Anthropic · closed · 2026-06
Claude Sonnet 5
not ranked n/a
the takeNear-Opus agentic muscle at half the Opus price, and the one I would hand most day-jobs. Off the board only because it is days old and reports vendor evals, not the standard suite. Watch this one.
Not ranked: Anthropic doesn't publish enough comparable public benchmarks to place it fairly. How ranking works.
MMLU-Pro
—
- Context
- 1M
- Input
- $3/M
- Output
- $15/M
- Speed
- 78.9 tok/s
- Modality
- text, image
- Near-Opus-4.8 agentic work at ~half price
- Highly autonomous, self-verifies
- Strong safety profile, 1M context
- Verbose, high output-token usage
- New tokenizer inflates token counts
- Too new for reliable independent scores